Pages:
Author

Topic: [Proposal] Mining Co-Op - page 4. (Read 11272 times)

full member
Activity: 140
Merit: 100
November 09, 2015, 07:33:28 PM


I want this project to have as little risk of scam as possible. That's why I've proposed multisig to collect funds, and after purchasing the miners never holding funds again. Using Kano's CKPool we can pay out mining revenue directly to multiple recipients.

Essentially, all we're doing is a group buy and group hosting. Simple!

Sign me up, asap. Tongue GND sounds good tbh, but we should get the future electricity price decrease in writing. Or some kind og guarantee Cheesy
hero member
Activity: 572
Merit: 506
November 09, 2015, 07:14:37 PM

Quote
How do you guys want to do this reward wise? Do you want to build a "company" of sorts, pool all resources and rewards to buy more hardware etc?
Or do you want to split all costs and rewards according to how many miners you "own"?

I want this project to have as little risk of scam as possible. That's why I've proposed multisig to collect funds, and after purchasing the miners never holding funds again. Using Kano's CKPool we can pay out mining revenue directly to multiple recipients.

Essentially, all we're doing is a group buy and group hosting. Simple!


Well, i think there is not the same buy a % of an SP-50 than 30 avalon-6. In the first case, we need to make a system to do multiple recipients. Buying 30 avalon-6 we can do it simple:

1.- We must buy entire miners (1, 2, 3, 5). With that we can hash with it where we want (and change it every time we want). And at the end, we can assign 1 miner to every one. And the miners configuration could be individually.
2.- We pay the miners as a group, so every one pay an escrow (the same) to make the deal with de seller. This escrow pay the seller the entire purchuase.
3.- We can buy the PSU as a group, and pay for them as a service. At the end, we can sell them as a group. As before we can pay an escrow to make the PSU deal.
4.- We pay the hosting as a group, so we do the same with the hosting fees. If we pay month by month, we can make the payments to the escrow, who make the last pay to the host.
5.- And finally, we can do more things if every one is happy...  Grin

What do you think?
hero member
Activity: 562
Merit: 506
We're going to need a bigger heatsink.
November 09, 2015, 06:23:26 PM
Okay, I finished my little script.

Here's the source if anyone is interested in taking a look.

Code:
# This is written in python 3
# Feel free to use this in your own projects

def truncate(f, n):
    '''Truncates/pads a float f to n decimal places without rounding'''
    s = '{}'.format(f)
    if 'e' in s or 'E' in s:
        return '{0:.{1}f}'.format(f, n)
    i, p, d = s.partition('.')
    return '.'.join([i, (d+'0'*n)[:n]])

###################
# HARCODED VALUES #
###################
#Cost in USD of miner (grand total)
minerCost = 1542.5

#Revenue of miner $/day
minerRevenue = 10.21

#Power cost of miner $/day
minerPowerCost = 1.67

#this block should be mined on November 27th
startBlock = 385389

#this is my projected difficulty on the 27th
startDiff = 66500000000
 
#this is the starting % increase every two weeks
expectedDiffIncrease = 4.0

#this is the % change in difficulty increase every two weeks
diffIncreaseDelta = -3.0

breakeven = False
revenue = 0
daysElapsed = 0 #in days
currentBlock = startBlock
currentDiff = startDiff
print("Simulation start: \nDifficulty is {}. \nProfit is ${} per day.\nExpected difficulty increase is {}%, with future increases changing by {}%.".format(truncate(currentDiff, 2), truncate(minerRevenue - minerPowerCost, 2), truncate(expectedDiffIncrease, 2), truncate(diffIncreaseDelta, 2)))
while True:
#1 loop = 1 block
if (currentBlock % 2016 == 0):
currentDiff *= (100 + expectedDiffIncrease)/100
minerRevenue *= (100 - expectedDiffIncrease)/100
print("Difficulty is now {} (+{}%). Profit is now ${} per day".format(truncate(currentDiff, 2), truncate(expectedDiffIncrease, 2), truncate(minerRevenue - minerPowerCost, 2)))
expectedDiffIncrease *= (100 + diffIncreaseDelta)/100

if (currentBlock == 420000):
minerRevenue /= 2
print("The block reward has halved. Profit is now ${} per day".format(truncate(minerRevenue - minerPowerCost, 2)))

if (minerRevenue - minerPowerCost <= 0):
print("Miner is no longer profitable. Ending simulation\n")
break

if (currentBlock - startBlock) % int((144*(100+expectedDiffIncrease)/100)//1) == 0: #calculates real time, note that this accounts for less time between blocks when diff spikes
daysElapsed += 1

if currentBlock == 450000:
print("End of simulation reached.\n")
break

if (revenue > minerCost and breakeven == False):
print("BREAK EVEN REACHED ON DAY {}!".format(daysElapsed))
breakeven = True #hell yes


revenue += minerRevenue/24/6 - minerPowerCost/24/6
currentBlock += 1

print("{} days have elapsed.".format(daysElapsed))
print("Your total revenue after hosting costs: {}".format(truncate(revenue,2)))
print("Your net gain/loss: ${}".format(truncate(revenue-minerCost,2)))

Here's the results, guessing an initial 4.0% increase every 2016 blocks. This assumes that the difficulty increase changes by -3% every adjustment.
Code:
Simulation start: 
Difficulty is 66500000000.00.
Profit is $8.54 per day.
Expected difficulty increase is 4.00%, with future increases changing by -3.00%.
Difficulty is now 69160000000.00 (+4.00%). Profit is now $8.13 per day
Difficulty is now 71843408000.00 (+3.88%). Profit is now $7.75 per day
Difficulty is now 74547306503.50 (+3.76%). Profit is now $7.39 per day
Difficulty is now 77268799058.20 (+3.65%). Profit is now $7.06 per day
Difficulty is now 80005019548.00 (+3.54%). Profit is now $6.75 per day
Difficulty is now 82753140848.50 (+3.43%). Profit is now $6.46 per day
Difficulty is now 85510382834.30 (+3.33%). Profit is now $6.19 per day
Difficulty is now 88274019729.60 (+3.23%). Profit is now $5.94 per day
Difficulty is now 91041386800.50 (+3.13%). Profit is now $5.70 per day
Difficulty is now 93809886395.30 (+3.04%). Profit is now $5.47 per day
Difficulty is now 96576993338.10 (+2.94%). Profit is now $5.26 per day
Difficulty is now 99340259691.70 (+2.86%). Profit is now $5.06 per day
Difficulty is now 102097318904.73 (+2.77%). Profit is now $4.88 per day
Difficulty is now 104845889362.98 (+2.69%). Profit is now $4.70 per day
Difficulty is now 107583777367.73 (+2.61%). Profit is now $4.53 per day
Difficulty is now 110308879565.73 (+2.53%). Profit is now $4.38 per day
Difficulty is now 113019184857.03 (+2.45%). Profit is now $4.23 per day
The block reward has halved. Profit is now $1.28 per day
Difficulty is now 115712775808.78 (+2.38%). Profit is now $1.21 per day
Difficulty is now 118387829603.62 (+2.31%). Profit is now $1.14 per day
Difficulty is now 121042618552.48 (+2.24%). Profit is now $1.08 per day
Difficulty is now 123675510201.35 (+2.17%). Profit is now $1.02 per day
Difficulty is now 126284967062.23 (+2.10%). Profit is now $0.96 per day
BREAK EVEN REACHED ON DAY 307!
Difficulty is now 128869545997.81 (+2.04%). Profit is now $0.91 per day
Difficulty is now 131427897289.49 (+1.98%). Profit is now $0.85 per day
Difficulty is now 133958763417.29 (+1.92%). Profit is now $0.81 per day
Difficulty is now 136460977579.81 (+1.86%). Profit is now $0.76 per day
Difficulty is now 138933461981.43 (+1.81%). Profit is now $0.72 per day
Difficulty is now 141375225912.73 (+1.75%). Profit is now $0.67 per day
Difficulty is now 143785363649.27 (+1.70%). Profit is now $0.63 per day
Difficulty is now 146163052192.63 (+1.65%). Profit is now $0.60 per day
Difficulty is now 148507548876.19 (+1.60%). Profit is now $0.56 per day
Difficulty is now 150818188857.25 (+1.55%). Profit is now $0.52 per day
End of simulation reached.

440 days have elapsed.
Your total revenue after hosting costs: 1641.04
Your net gain/loss: $98.54

My difficulty estimate is not very optimistic, although it's far from a worst case. We could do far better, or we could stand to lose money overall. And remember, this assumes a static BTC price, it could go anywhere.

Quote
How do you guys want to do this reward wise? Do you want to build a "company" of sorts, pool all resources and rewards to buy more hardware etc?
Or do you want to split all costs and rewards according to how many miners you "own"?

I want this project to have as little risk of scam as possible. That's why I've proposed multisig to collect funds, and after purchasing the miners never holding funds again. Using Kano's CKPool we can pay out mining revenue directly to multiple recipients.

Essentially, all we're doing is a group buy and group hosting. Simple!
full member
Activity: 140
Merit: 100
November 09, 2015, 05:21:16 PM
Those figures don't include the halving, shipping ($20-50), import duties($67.5), or PSUs. For PSU cost, bulk buying the "royale with cheese" package from finksy could probably get down to $200/unit shipped (maybe lower), and that package can power exactly 2.5 Avalon6 units. Call that 80 for the PSU.

Our total cost per miner and PSU @ 50 units purchased would be $1542.5 USD a unit.

I'm going to write a quick python script to crunch these numbers.

Knew I missed something ^^. Get back to us asap with your script Cheesy

How do you guys want to do this reward wise? Do you want to build a "company" of sorts, pool all resources and rewards to buy more hardware etc?
Or do you want to split all costs and rewards according to how many miners you "own"?

Personally I want to grow this thing into something big, by pooling resources. I'm a firm believer in bitcoin. I'm not saying you can't combine the two tho, just want to know what you guys think.

Phil: I also believe the price would be higher at July, and much higher in the future.

I'm ready right now to contribute with 5 Avalon's at that price.

legendary
Activity: 4116
Merit: 7849
'The right to privacy matters'
November 09, 2015, 05:01:09 PM
I don't see the diff number you used  did you punch in 2 % per jump 3% per jump

I see the monthly diff goes up but how did you get it?

Never mind I found the number 5% per month.  you kept coins flat at 380.

and you have break even around the ½ ing in July

I think that could all happen.
 I actually think price will do better then 380 from now to July.

ROI for 1 unit:


ROI for 30 units:


What do you think about these numbers? Ofc there is some hidden costs.
hero member
Activity: 562
Merit: 506
We're going to need a bigger heatsink.
November 09, 2015, 04:54:38 PM
Those figures don't include the halving, shipping ($20-50), import duties($67.5), or PSUs. For PSU cost, bulk buying the "royale with cheese" package from finksy could probably get down to $200/unit shipped (maybe lower), and that package can power exactly 2.5 Avalon6 units. Call that 80 for the PSU.

Our total cost per miner and PSU @ 50 units purchased would be $1542.5 USD a unit.

I'm going to write a quick python script to crunch these numbers.
full member
Activity: 140
Merit: 100
November 09, 2015, 04:29:24 PM
ROI for 1 unit:


ROI for 30 units:


What do you think about these numbers? Ofc there is some hidden costs.
hero member
Activity: 562
Merit: 506
We're going to need a bigger heatsink.
November 09, 2015, 03:52:29 PM
I know a little about GND  I know nothing about CryptoBoreas

So what is the link for CB?

What are power costs for GND vs CB?

GND is in Canada where is CB?
CryptoBoreas is the rebranded name of Allinvain's hosting service. Link: https://bitcointalksearch.org/topic/ann-cryptoboreas-cheap-miner-hosting-1170850

CB advertised $65 CAD/KW ($49 USD) for a 6 month term, great north data advertises $60/KW (USD), but can go a bit lower on a 6mo term and a large amount of hardware. Great North Data will have much lower rates in a few months, but right now CB wins.

The datacenter for CryptoBoreas is in Labrador, Canada, and although they don't specify what city they are in it's safe to assume they are based out of Goose Bay. No other city in the province has the right combination of power costs and reliable internet access.
legendary
Activity: 4116
Merit: 7849
'The right to privacy matters'
November 09, 2015, 03:19:02 PM
WOOOHOO! I just got off the phone with blockC, and I can post something very, very rare to this thread: ACTUAL NUMBERS!
I made sure they were OK with me sharing this information, so put your lawsuits/pitchforks away.  Smiley

Unit Cost:
For a group buy of moderate size they can do $1350/unit, raspberry pis included. We may get a small additional discount if we get a buttload of them (50+).

Shipping Cost:
If we get less than 50 units it will cost $50/unit (via UPS, ew),
If we get over 50 units they can ship via air freight. That's ~$20 a unit, and can get lower if we buy more. If we got 100 units (I know, it's a big ask), I've been told it goes down to $10 per unit.

This beats the S7 price per TH, arrives earlier, and has practically the same efficiency (0.28J/GH for the S7, 0.29J/GH for the Avalon6). I've been told you can get the Avalon6 down to 0.28 if you use very efficient PSUs, which we are planning on getting with them.

Timeframe:
These are batch 2 avalon6 miners. They are scheduled to arrive at BlockC November 20th, and should arrive at GND/CryptoBoreas about a week after that. If we don't raise funds in time, no big deal. There will be a batch 3, 4, etc. Of course, I'd prefer sooner rather than later.

Undervolting:
The Avalon6 has a string design, so it will be hard to undervolt. Much like the S5/S7, you'll need to actually adjust the input voltage into the miner. It's not very practical, although blockC is planning on making a DC/DC converter in the future that will do the trick.

Note that the S7 does not have a 0.25J/GH efficiency as many people believe. If you look at the specs page for batch 6, it says 0.25J/GH + 10%.

I know a little about GND  I know nothing about CryptoBoreas

So what is the link for CB?

What are power costs for GND vs CB?

GND is in Canada where is CB?

_________________________________________________________________________
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@



The theory of undervolting an avalon 6 is yet to be proven or tested. 

In a 5, 6 ,7, 8  cent a kwatt  location it is not much of an issue.

That is more for someone like me my winter in house rate = 10 cents my summer in house rate = 17 cents.

Running 1 avalon 6 at 10 volts for the summer is something I may do.
hero member
Activity: 562
Merit: 506
We're going to need a bigger heatsink.
November 09, 2015, 03:12:01 PM
I agree that right now cryptoboreas is the best option pricewise right now. I'm pretty sure GND hasn't finished their datacenter expansion yet, which means their rates are still going to be quite a bit higher than CB.

On the other hand, great north data has a 24/7 onsite tech, something cryptoboreas doesn't have. If there's a problem, its likely that we will get our miners back up and running faster if we go with GND.

It really depends if you guys think its worth it or not.
hero member
Activity: 572
Merit: 506
November 09, 2015, 03:04:16 PM
WOOOHOO! I just got off the phone with blockC, and I can post something very, very rare to this thread: ACTUAL NUMBERS!
I made sure they were OK with me sharing this information, so put your lawsuits/pitchforks away.  Smiley

Unit Cost:
For a group buy of moderate size they can do $1350/unit, raspberry pis included. We may get a small additional discount if we get a buttload of them (50+).

Shipping Cost:
If we get less than 50 units it will cost $50/unit (via UPS, ew),
If we get over 50 units they can ship via air freight. That's ~$20 a unit, and can get lower if we buy more. If we got 100 units (I know, it's a big ask), I've been told it goes down to $10 per unit.

This beats the S7 price per TH, arrives earlier, and has practically the same efficiency (0.28J/GH for the S7, 0.29J/GH for the Avalon6). I've been told you can get the Avalon6 down to 0.28 if you use very efficient PSUs, which we are planning on getting with them.

Timeframe:
These are batch 2 avalon6 miners. They are scheduled to arrive at BlockC November 20th, and should arrive at GND/CryptoBoreas about a week after that. If we don't raise funds in time, no big deal. There will be a batch 3, 4, etc. Of course, I'd prefer sooner rather than later.

Undervolting:
The Avalon6 has a string design, so it will be hard to undervolt. Much like the S5/S7, you'll need to actually adjust the input voltage into the miner. It's not very practical, although blockC is planning on making a DC/DC converter in the future that will do the trick.

Note that the S7 does not have a 0.25J/GH efficiency as many people believe. If you look at the specs page for batch 6, it says 0.25J/GH + 10%.

Hi, my vote is for Cryptoboreas, i have with them one S7, one SP20 and two S5 (they are setting up last two units now).  Grin
hero member
Activity: 562
Merit: 506
We're going to need a bigger heatsink.
November 09, 2015, 02:40:11 PM
WOOOHOO! I just got off the phone with blockC, and I can post something very, very rare to this thread: ACTUAL NUMBERS!
I made sure they were OK with me sharing this information, so put your lawsuits/pitchforks away.  Smiley

Unit Cost:
For a group buy of moderate size they can do $1350/unit, raspberry pis included. We may get a small additional discount if we get a buttload of them (50+).

Shipping Cost:
If we get less than 50 units it will cost $50/unit (via UPS, ew),
If we get over 50 units they can ship via air freight. That's ~$20 a unit, and can get lower if we buy more. If we got 100 units (I know, it's a big ask), I've been told it goes down to $10 per unit.

This beats the S7 price per TH, arrives earlier, and has practically the same efficiency (0.28J/GH for the S7, 0.29J/GH for the Avalon6). I've been told you can get the Avalon6 down to 0.28 if you use very efficient PSUs, which we are planning on getting with them.

Timeframe:
These are batch 2 avalon6 miners. They are scheduled to arrive at BlockC November 20th, and should arrive at GND/CryptoBoreas about a week after that. If we don't raise funds in time, no big deal. There will be a batch 3, 4, etc. Of course, I'd prefer sooner rather than later.

Undervolting:
The Avalon6 has a string design, so it will be hard to undervolt. Much like the S5/S7, you'll need to actually adjust the input voltage into the miner. It's not very practical, although blockC is planning on making a DC/DC converter in the future that will do the trick.

Note that the S7 does not have a 0.25J/GH efficiency as many people believe. If you look at the specs page for batch 6, it says 0.25J/GH + 10%.
legendary
Activity: 1484
Merit: 1004
November 09, 2015, 02:13:47 PM
I do agree with the choice of the PSU.

Incorporation could also be a good idea.  Grin
hero member
Activity: 562
Merit: 506
We're going to need a bigger heatsink.
November 09, 2015, 01:25:16 PM
Updates:
  • Ehash doesn't seem to be selling the Avalon6, and they redirected me to BlockC. Those guys got back to me with a quote! I'm not going to post it in the thread (yet!) but it wasn't a bad offer.
  • I've sent bitmain a very similar email, and have yet to hear back from them.
  • SP-Tech got back to my second email. No change in status since last time: No additional orders, no pricing, no MOQ, and no schedule for future production. The SP50 won't be an option for a while.

I've contacted Avalon for a quote on 10, 30, 50 and 100 units. I guess we'll wait and see.  Smiley

Awesome!

But, how should we do this exactly? I guess we need to make some kind of corp/org?

I have no experience with incorporation and all my knowledge comes from a few hours of research. I remember a user called Dr Charles earlier in this thread recommended incorporating as well, and he said he had a ton of experience: I've sent him a PM about it. Hopefully he gets back to me!

As for PSUs, I think we should go for Finksy's 2880W units: https://bitcointalksearch.org/topic/ibm-2880w-psubreakout-boards-new-package-deals-for-t9s9-a7-966135
full member
Activity: 140
Merit: 100
November 09, 2015, 12:25:14 PM
I've contacted Avalon for a quote on 10, 30, 50 and 100 units. I guess we'll wait and see.  Smiley

Awesome!

But, how should we do this exactly? I guess we need to make some kind of corp/org?
legendary
Activity: 1456
Merit: 1000
November 09, 2015, 11:24:04 AM
avalon 6 Shocked Shocked Shocked Shocked Shocked Shocked Shocked Shocked

better batch 6 of s7!

All depends on the underclock which I'm hoping we will see anyday.  If the Avalon 6 has the underclock like 4.1 you will want these to be able to underclock.

But if it's underclock is not like 4.1 then it's not as big of deal.  So we should see pretty soon hopefully on underclock.
hero member
Activity: 742
Merit: 500
November 09, 2015, 07:26:03 AM
avalon 6 Shocked Shocked Shocked Shocked Shocked Shocked Shocked Shocked

better batch 6 of s7!
hero member
Activity: 562
Merit: 506
We're going to need a bigger heatsink.
November 08, 2015, 09:41:27 PM
I've contacted Avalon for a quote on 10, 30, 50 and 100 units. I guess we'll wait and see.  Smiley
hero member
Activity: 572
Merit: 506
November 08, 2015, 05:04:35 PM
So, get started with 25x S7? 30 kW. At least we got the structure up and running for future investments!

I think the AVALON 6 is a better option if it can be underclocked... Wink

Something like 30 or 35 avalon6 to start?  Roll Eyes
hero member
Activity: 562
Merit: 506
We're going to need a bigger heatsink.
November 07, 2015, 01:00:12 PM
So, get started with 25x S7? 30 kW. At least we got the structure up and running for future investments!

That's an interesting point. We could start out with a small buy of hardware and ramp up from there. On the other hand, we likely won't get any volume discounts for a 30KW purchase, so there's probably going to be little interest in the small buy.
Pages:
Jump to: